Understanding Salesforce Data Cloud and AgentForce Costing: A Financial Services Case Study

Introduction

In today's digital landscape, financial institutions need robust systems to handle vast amounts of data while providing intelligent customer service. This article I am trying to work through a very simple hypothetical example of a financial services organization (picked up from recent experience:)), where Salesforce Data Cloud and AgentForce work together.

In this example, we will only be considering one data source (let’s say Salesforce). And we will only look at costing for Ingestion, Storage and Segmentation primarily. Along with it we will explore costing around AgentForce base as well as conversation costs. Our team at CloudColon is working on a tool to be able to estimate this cost more accurately, but as of now these are representational examples and this is more of a thought exercise I wanted to do. Hence, I don’t claim this calculation to be accurate.

Let’s consider this financial services organization is processing 75,000 loan applications daily. Further volume assumptions below:

Case Study: FinServ Lending Corp

Data Volume Overview

• Daily loan applications: 75,000
• Associated records per application (average):

1.?????? Credit check results: 1 record

2.?????? Income verification documents: 3 records

3.?????? Customer interaction logs: 2 records

4.?????? Collection status updates: 1 record

5.?????? Service requests: 0.5 records (not all applications generate service requests)

Total daily records are approx.: ~562,500 (75,000 × 7.5 associated records)

Data Cloud Pricing Breakdown

1. Data Ingestion Costs

Monthly Data Volume

• Primary records: 1.5M (75,000 × 20 business days)
• Associated records: 11.25M
• Total monthly records: 12.75M

Pricing Tiers (representative example)

• First 1M profiles: $0.0075 per profile
• 1M-10M profiles: $0.005 per profile
• Above 10M profiles: $0.003 per profile

Monthly Data Ingestion Cost:

• First 1M: $7,500
• Next 9M: $45,000
• Remaining 2.75M: $8,250 Total: $60,750/month

2. Data Storage Costs

Assuming 6-month retention:

• Total stored profiles: 76.5M (12.75M × 6 months)
• Storage cost: $0.001 per profile per month
• Monthly storage cost: $76,500

3. Calculated Insights and ML Model Costs

Basic Segmentation

• Customer risk scoring
• Payment probability prediction
• Collection priority calculation Base cost: $2,000/month per model

For 3 basic ML models: Total ML cost: $6,000/month

?

AgentForce Cost Analysis

Agent Configuration

• Number of cases per day: 3,750 (assumption: 5% of daily applications for inquiries like Loan agreement document, current stage of loan, EMI next due date etc)
• Average conversation duration: 8 minutes (assumption)
• Operating hours: 12 hours/day
• Required concurrent agents: 42

Agent Pricing Components

1. Base License Cost: $50 per agent per month Monthly cost: $2,100 (42 agents)
2. Conversation Processing: $0.03 per minute Monthly minutes: 600,000 (3,750 cases × 8 minutes × 20 days) Monthly cost: $18,000 [This is with an assumption that it is still going to take 8 mins however it should be way lesser than 8 mins. That is where efficiency gain will exist]
3. Data Cloud Integration: $0.01 per profile lookup Monthly lookups: 75,000 (3,750 cases × 20 days) Monthly cost: $750

Total Monthly AgentForce Cost: $20,850

Visual Cost Breakdown

Here is a visualization of the monthly costs:

Cost Optimization Recommendations

We can optimize the cost further by following pointers:

1. Data Ingestion Optimization Implement data filtering to reduce unnecessary record creation Use batch processing for non-urgent updates Estimated savings: 15-20%
2. Storage Optimization Implement tiered storage strategy Archive older records to cheaper storage Potential savings: 25-30%
3. Agent Utilization Implement intelligent routing Use pre-built response templates Expected efficiency gain: 20%

?

PS: We are working on ROI calculator as well but doing some simple math, cost for 42 agents to handle 3500 cases a day vs AgentForce cost will have a significant ROI.